Immunoregulatory and
anti-inflammatory effects of
n
-3
polyunsaturated fatty acids
Division of Human Nutrition, School of Biological Sciences, University of Southampton, Southampton, UKP.C. Calder
Abstract
1. Fish oils are rich in the long-chain n-3 polyunsaturated fatty acids (PUFAs), eicosapentaenoic (20:5n-3) and docosahexaenoic (22:6n-3) acids. Linseed oil and green plant tissues are rich in the precursor fatty acid, α-linolenic acid (18:33). Most vegetable oils are rich in the n-6 PUFA linoleic acid (18:2n-n-6), the precursor of arachidonic acid (20:4n-6).
2. Arachidonic acid-derived eicosanoids such as prostaglandin E2 are pro-inflammatory and regulate the functions of cells of the immune system. Consumption of fish oils leads to replacement of arachidonic acid in cell membranes by eicosapentaenoic acid. This changes the amount and alters the balance of eicosanoids produced.
3. Consumption of fish oils diminishes lymphocyte proliferation, T-cell-mediated cytotoxicity, natural killer cell activity, macrophage-mediated cytotoxicity, monocyte and neutrophil chemotaxis, major histocompatibility class II expression and antigen presentation, pro-duction of pro-inflammatory cytokines (interleukins 1 and 6, tumour necrosis factor) and adhesion molecule expression.
4. Feeding laboratory animals fish oil reduces acute and chronic inflammatory responses, improves survival to endotoxin and in mod-els of autoimmunity and prolongs the survival of grafted organs. 5. Feeding fish oil reduces cell-mediated immune responses. 6. Fish oil supplementation may be clinically useful in acute and chronic inflammatory conditions and following transplantation. 7. n-3 PUFAs may exert their effects by modulating signal transduc-tion and/or gene expression within inflammatory and immune cells.
Correspondence
P.C. Calder
Division of Human Nutrition School of Biological Sciences University of Southampton Bassett Crescent East Southampton SO16 7PX UK
Fax: 44-1703-594383
Presented at the XII Annual Meeting of the Federação de Sociedades de Biologia Experimental, Caxambu, MG, Brasil, August 27-30, 1997. Received August 21, 1997 Accepted August 25, 1997 Key words •Lymphocyte •Macrophage •Immune system •Inflammation •Fish oil
•Polyunsaturated fatty acid •Cytokine
•Eicosanoid
Biosynthesis and sources of different fatty acids
All mammals can synthesise fatty acids
de novo from acetyl coenzyme A. The end
product of the fatty acid synthetase enzyme is palmitic acid (16:0), which in turn can be elongated to stearic acid (18:0). There is little need for the synthesis of saturated fatty acids in Western man, since the diet
nor-mally supplies adequate amounts. However, cell membranes require unsaturated fatty acids to maintain their structure, fluidity and function. Therefore, a mechanism for the introduction of double bonds (i.e., desatura-tion) exists. The introduction of a single double bond between carbon atoms 9 and 10 is catalysed by the enzyme ∆9-desaturase, which is universally present in both plants and animals. This enzyme catalyses the
coversion of stearic acid to oleic acid (18:1 n-9). Plants, unlike animals, can insert addi-tional double bonds into oleic acid between the existing double bond at the 9 position and the methyl terminus of the carbon chain; a ∆12-desaturase converts oleic acid into li-noleic acid (18:2 n-6) while a ∆15-desaturase converts linoleic acid into α-linolenic acid (18:3 n-3). Since animal tissues are unable to synthesise linoleic and α-linolenic acids, these fatty acids must be consumed in the diet and so are termed essential fatty acids. Using the pathway outlined in Figure 1, ani-mal cells can convert dietary α-linolenic acid into eicosapentaenoic acid (20:5 n-3) and docosahexaenoic acid (22:6 n-3); by a similar series of reactions dietary linoleic acid is converted via γ-linolenic (18:3 n-6) and dihomo-γ-linolenic (20:3 n-6) acids to arachidonic acid (20:4 n-6). The n-9, n-6 and
n-3 families of polyunsaturated fatty acids
(PUFAs) are not metabolically interconvert-ible in mammals. Many marine plants, espe-cially the unicellular algae in phytoplankton, also carry out chain elongation and further desaturation of α-linolenic acid to yield the long-chain n-3 PUFAs eicosapentaenoic and docosahexaenoic acids. It is the formation of these long chain n-3 PUFAs by marine algae and their transfer through the food chain to fish that accounts for their abundance in some marine fish oils.
Influence of dietary n-3 PUFAs on the functions of cells of the immune system
Many studies have investigated the ef-fects of the amount and type of fat in the diet upon immune cell functions, particularly lym-phocyte proliferation in response to mito-gens. These studies have been reviewed sev-eral times in recent years (1-9); the effects of
n-3 PUFAs are the most well documented
and will be summarised here.
Lymphocyte proliferation
Lymphocyte proliferation is usually meas-ured as the incorporation of radioactively labelled precursors (e.g., thymidine) into DNA. A suitable stimulus (termed a mito-gen) to activate the lymphocytes is required; mitogens used most frequently are con-canavalin A (Con A) and phytohaemaggluti-nin (PHA) which stimulate T lymphocytes specifically, bacterial lipopolysaccharide (LPS) which stimulates B lymphocytes and pokeweed mitogen (PWM) which stimulates a population of both T and B lymphocytes. Monoclonal antibodies to lymphocyte sur-face structures (e.g., CD3) can also be used to stimulate proliferation.
Feeding animals (rats, mice, chickens, rabbits) diets containing high levels (70 to 200 g/kg) of fish oil (rich in eiocosapen-taenoic and docosahexaenoic acids) has been shown to result in suppressed proliferation of lymphocytes stimulated with T- or B-cell mitogens compared with feeding diets rich in other fats such as lard, coconut oil, corn oil, safflower oil or linseed oil (10-15). Re-cently, it was reported that eicosapentaenoic and docosahexaenoic acids are equipotent in reducing murine spleen lymphocyte prolif-eration (16). Diets containing large amounts of linseed oil (rich in α-linolenic acid) de-crease lymphocyte proliferation compared with saturated fatty acid- or n-6 PUFA-rich diets (12,17,18). Recently, it was shown that increasing the amount of α-linolenic acid in the rat diet at the expense of linoleic acid decreases spleen lymphocyte proliferation (19).
Sometimes, the effects of n-3 PUFA-rich diets on lymphocyte proliferation were dem-onstrated when the cells were cultured in autologous serum, but were lost if the cells were cultured in foetal calf serum (11-13). Culture of cells in foetal calf serum may explain the lack of effect upon spleen
lym-phocyte proliferation of feeding fish and linseed oils to mice reported in some studies (20,21). It has been shown that the changes in lymphocyte fatty acid composition in-duced by dietary manipulations are better maintained if the cells are cultured in autolo-gous rather than foetal calf serum (22).
Twelve weeks of supplementation of the
diets of healthy women (51 to 68 years of age) with encapsulated n-3 PUFAs (approxi-mately 2.4 g/day) resulted in a lowered mito-genic response of peripheral blood mono-nuclear cells (PBMNCs) to PHA (23). More recently, a decreased response of PBMNCs to Con A and PHA following supplementa-tion of the diet of volunteers on a low fat, low
stearic acid 18:0 ∆9 α-linolenic acid 18:3n-3 ∆15 (plants) linoleic acid 18:2n-6 ∆12 (plants) oleic acid 18:1n-9 ∆6 ∆6 18:4n-3 γ-linoleic acid 18:3n-6 Further desaturation and elongation elongase elongase 20:4n-3
dihomo-γ-linoleic acid 20:3n-6 ∆5 ∆5 eicosapentaenoic acid 20:5n-3 arachidonic acid 20:4n-6 elongase elongase 22:5n-3 22:4n-6 elongase elongase 24:5n-3 24:4n-6 ∆6 ∆6 24:6n-3 24:5n-6 β-oxidation β-oxidation docosahexaenoic acid 22:6n-3 22:5n-6
Figure 1 - PUFA metabolism. ∆5,
∆6, ∆9, ∆12 and ∆15 indicate desaturase enzymes.
cholesterol diet with 1.23 g n-3 PUFAs/day was reported (24), while 18 g of fish oil/day (approximately 6 g n-3 PUFAs/day) for 6 weeks resulted in lowered PHA-stimulated proliferation of PBMNCs 10 weeks after supplementation had ended (but not at the end of the supplementation period) (25). Cytotoxic T lymphocyte-mediated cytolysis
The cytotoxic T lymphocyte activity of spleen lymphocytes was reported to be lower after feeding mice 100 g/kg fish oil for up to 10 weeks than after feeding 100 g/kg linseed oil (26). Feeding chickens diets containing 70 g/kg fish or linseed oil significantly re-duced spleen cytotoxic T-cell activity com-pared with diets containing 70 g/kg lard or corn oil (27).
Natural killer cell-mediated cytolysis
Feeding mice diets containing 100 g/kg fish oil caused a decrease in spleen natural killer (NK) cell activity compared with feed-ing chow or 100 g/kg corn oil (28,29). In the study of Berger et al.(21) female mice were fed for 5 months on diets containing 100 g/ kg olive, safflower, linseed or fish oil and the spleen NK cell activity of the pups was determined before they were weaned; the activity was lower in the fish oil group than in the safflower or olive oil groups. Yaqoob et al. (30) fed weanling rats for 10 weeks on a low fat diet or on diets containing 200 g/kg hydrogenated coconut, olive, safflower, evening primrose or fish oil before measur-ing spleen NK cell activity. It was found that feeding each of the high fat diets resulted in lower NK cell activity compared with feed-ing the low fat diet; feedfeed-ing the fish oil diet resulted in the lowest activity. Similar re-sults have been found in more mature rats fed on these diets for 12 weeks (15). It was reported that a 200 g/kg linseed oil diet decreased rat spleen lymphocyte NK cell activity compared with feeding a 200 g/kg
sunflower oil diet(18). More recently, it was shown that increasing the amount of α -lino-lenic acid in the rat diet at the expense of linoleic acid decreases spleen NK cell activ-ity (19).
No studies have investigated the effect of dietary lipids on human NK cell activity, although it was reported that intravenous injection of a triacylglycerol-containing eicosapentaenoic acid into healthy human volunteers results in suppression of periph-eral blood NK cell activity 24 h later (31). Macrophage-mediated cytotoxicity
Dietary fish oil (100 g/kg) has been re-ported to significantly suppress lysis of tar-get tumour cells by mouse-elicited perito-neal macrophages (32-35). The target cell lines used in these studies are sensitive to killing by tumour necrosis factor (TNF)-α (L929 cells; 33,34) or nitric oxide (P815 cells; 32,35). Thus, the suppressed macro-phage-mediated cytolysis observed after fish oil feeding suggests that fish oil reduces the production of nitric oxide and TNF (see later).
Neutrophil and monocyte chemotaxis Chemotaxis of human peripheral blood neutrophils and monocytes towards a variety of chemoattractants including leukotriene (LT) B4, platelet-activating factor, formyl-methionyl-leucyl-phenylalanine and autolo-gous serum is suppressed following the sup-plementation of the human diet with n-3 PUFAs (36-40).
MHC expression and antigen presentation Inclusion of n-3 PUFAs in the diet of mice or rats results in a diminished percent-age of peritoneal exudate cells bearing the major histocompatibility class (MHC) II an-tigens on their surface (41-44). The level of MHC II expression on positive cells is also
suppressed by fish oil feeding (43). In accor-dance with these animal studies, supplemen-tation of the diet of human volunteers with
n-3 PUFAs (approximately 1.56 g/day) for 3
weeks resulted in a decreased level of MHC II (HLA-DP, -DQ and -DR) expression on the surface of peripheral blood monocytes (45). These observations suggest that diets rich in n-3 PUFAs will result in diminished antigen presentation. Indeed, feeding mice the ethyl ester of eicosapentaenoic acid for a period of 4 to 5 weeks resulted in diminished presentation of antigen (keyhole limpet haemocyanin; KLH) by spleen cells ex vivo (46). Dendritic cells are the key antigen-presenting cells in vivo. We have recently observed that, compared with a low fat diet or a diet containing 200 g/kg safflower oil, feeding rats a diet containing 200 g/kg fish oil significantly diminishes MHC II expres-sion on dendritic cells and ex vivo antigen (KLH) presentation by dendritic cells (ob-tained by cannulation of the thoracic duct) to KLH-sensitised spleen lymphocytes (47).
Influence of dietary n-3 PUFAs on the interactions between cells of the immune system
Communication between cells of the im-mune system is achieved by virtue of the production of chemical mediators (e.g., eicosanoids, cytokines, nitric oxide (NO)) and by direct cell-to-cell contact mediated by adhesion molecules. n-3 PUFAs have been found to influence each of these com-munication links.
n-3 PUFAs and eicosanoid production Eicosanoids are a family of oxygenated derivatives of dihomo-γ-linolenic, arachi-donic and eicosapentaenoic acids. Eicosa-noids include prostaglandins (PGs), throm-boxanes (TXs), LTs, lipoxins (LXs), hydro-peroxyeicosatetraenoic acids (HPETEs) and hydroxyeicosatetraenoic acids (HETEs). In
most conditions the principal precursor for these compounds is arachidonic acid. The precursor PUFA is released from membrane phosphatidylcholine by the action of phos-pholipase A2 or from membrane phosphati-dylinositol-4,5-bisphosphate (PIP2) by the actions of phospholipase C (PLC) and a diacylglycerol (DAG) lipase. The pathways of eicosanoid synthesis begin with cyclo-oxygenase, which yields the PGs and TXs, or with the 5-, 12- or 15-lipoxygenases, which yield the LTs, HPETEs, HETEs and LXs (Figure 2). The amounts and types of eicosanoids synthesised are determined by the availability of arachidonic acid, by the activities of phospholipase A2 and PLC, by the activities of cyclooxygenase and the li-poxygenases, by the cell type and by the nature of the stimulus.
Cells of the immune system are an impor-tant source of eicosanoids and they are sub-ject to their regulatory effects (48-51); the most well-documented effects are those of PGE2. In vivo, PGs are involved in modulat-ing the intensity and duration of inflamma-tory and immune responses; PGE2 has a
Phospholipid Phospholipase A2 Arachidonic acid (20:4n-6) Eicosapentaenoic acid (20:5n-3) Cyclooxygenase PGI2 TXA2 TXB2 PGD2 PGE2 PGF2 Lipoxygenases LTA4 LTB4 LTC4 LTD4 LTE4 5-HPETE 5-HETE 15-HPETE 15-HETE 12-HPETE 12-HETE Cyclooxygenase PGI3 TXA3 TXB3 PGD3 PGE3 PGF3 Lipoxygenases LTA5 LTB5 LTC5 LTD5 LTE5 5-HEPE
Figure 2 - Synthesis of eicosanoids from arachidonic and eicosapentaenoic acids. HEPE indicates hydroxyeicosapentaenoic acid; see text for explanation of other abbreviations.
number of pro-inflammatory effects includ-ing induction of fever and erythema, in-creased vascular permeability and vasodila-tion and enhancement of pain and oedema caused by other agents such as bradykinin and histamine. PGE2 also regulates the pro-duction of both monocyte/macrophage- and lymphocyte-derived cytokines (see 52). In chronic inflammatory conditions increased rates of PGE2 production are found, and elevated PGE2 production has been observed in patients suffering from infections. LTs have chemotactic properties and are involved in the regulation of inflammatory and im-mune processes; 4-series LTs regulate cyto-kine production.
The n-3 PUFAs, eicosapentaenoic and docosahexaenoic acids, competitively inhibit the oxygenation of arachidonic acid by cy-clooxygenase. In addition, eicosapentaenoic acid (but not docosahexaenoic acid) is able to act as a substrate for both cyclooxygenase and 5-lipoxygenase. Thus, ingestion of fish oils which contain n-3 PUFAs will result in a decrease in membrane arachidonic acid levels and a concomitant decrease in the capacity to synthesise eicosanoids from arachidonic acid (e.g., 53); eicosapentaenoic acid gives rise to the 3-series PGs and TXs and the 5-series LTs (Figure 2). The eicosan-oids produced from eicosapentaenoic acid do not always have the same biological prop-erties as the analogues produced from arachi-donic acid.
n-3 PUFAs and cytokine production
Since cytokine production is regulated by eicosanoids (see 52) and since dietary lipids affect eicosanoid production, it might be expected that dietary lipids, especially those containing n-3 PUFAs, will affect cytokine production. The effects of n-3 PUFAs on cytokine production have been reviewed several times in recent years (7,54-59).
Animal studies. All studies which have
used murine resident peritoneal macrophag-es (60-65), one study using rat rmacrophag-esident al-veolar macrophages (66) and one study us-ing rat resident peritoneal macrophages (67) report an enhancing effect of n-3 PUFAs upon TNF production; only one study which used rat resident peritoneal macrophages has reported reduced TNF production following fish oil feeding (68). Three studies report that n-3 PUFA-rich diets do not affect TNF production by CFA-elicited peritoneal mac-rophages from rats or mice (62,63,67). The effect of dietary n-3 PUFAs upon TNF pro-duction by thioglycollate-elicited peritoneal macrophages is unclear, with studies report-ing no effect (35,63,69,70), reduction (34, 53,71) or enhancement (35,69). Comparison of the outcome of these studies is compli-cated by the different procedures used for ex
vivo culture of the cells. That such details are
important in determining outcome is shown by the studies of Erickson’s group (35,69) in which there was no effect of diet upon TNF production when thioglycollate-elicited mac-rophages were stimulated with LPS for 1 or 8 h but if they were stimulated for 24 h there was increased TNF production by cells from fish oil-fed mice. The only animal study which has investigated TNF production by PBMNCs showed decreased production fol-lowing the infusion of a 10% (v/v) fish oil emulsion (72); this is an interesting observa-tion since it agrees with the findings of a number of studies using human PBMNCs (see below). In addition to studies measuring TNF production ex vivo, Black and Kinsella (33) and Renier et al. (34) showed that feed-ing mice n-3 PUFA-rich diets resulted in reduced ability of elicited peritoneal macro-phages to kill L929 cells; L929 cells are killed by TNF and so the reduced cytotoxic-ity of macrophages towards these cells sug-gests a reduced ability to produce TNF. Two animal studies have investigated the effects of dietary lipids upon circulating TNF levels which would perhaps reflect in vivo produc-tion of the cytokine. Watanabe et al. (63)
found that TNF levels were significantly higher in the plasma of LPS-injected mice fed a diet containing perilla oil than in the plasma of those fed a safflower oil-rich diet. Similarly, Chang et al. (65) reported higher serum TNF levels 1 and 1.5 h following intraperitoneal injection of LPS into fish oil-fed mice compared with those oil-fed coconut or corn oil.
All studies which have used thioglycol-late-elicited peritoneal macrophages and the only study to use Kupffer cells report that dietary fish oil results in decreased ex vivo production of IL-1 (34,53,70,71,73). In con-trast, two studies have reported that fish oil enhances IL-1 production by murine resi-dent macrophages (60,64). In addition, Ertel et al. (74) showed that the reduction in ex
vivo IL-1 production by resident peritoneal
macrophages which accompanies haemor-rhagic shock in corn or safflower oil-fed mice was prevented by fish oil feeding. This study also showed no difference in IL-1 production by resident peritoneal macrophag-es taken from sham-operated mice fed thmacrophag-ese three diets (74).
There are no studies reporting the effect of dietary fatty acids upon IL-6 production by resident peritoneal macrophages. One study using murine thioglycollate-elicited peritoneal macrophages showed a signifi-cant reduction in LPS-stimulated IL-6 duction following fish oil feeding (53); pro-duction following stimulation of the cells with TNF was also significantly reduced (Yaqoob P and Calder PC, unpublished ob-servations). Rat PBMNCs showed reduced IL-6 production following fish oil infusion for 4 days (72).
In contrast to the large number of studies in animals of the effects of dietary lipids, especially fish oil, upon the ex vivo produc-tion of macrophage-derived cytokines (see above), there have been relatively few stud-ies on lymphocyte-derived cytokines. De-creased IL-2 production by alveolar lympho-cytes from pigs fed diets containing fish or
linseed oil was reported (75). Recently, it was observed that feeding mice a diet con-taining 10 g/kg of ethyl esters of eicosapen-taenoic or docosahexaenoic acid for 10 days significantly decreased ex vivo IL-2 produc-tion by spleen lymphocytes stimulated with Con A; both n-3 PUFAs were equally effec-tive (16). In another study weanling mice were fed for 8 weeks on a low fat diet or on diets containing 200 g/kg hydrogenated co-conut, olive, safflower or fish oil; the spleen lymphocytes were subsequently stimulated with Con A (14). The concentration of IL-2 was higher in the medium of spleen lympho-cytes from mice fed olive or safflower oil than in the medium of cells from mice fed the low fat diet or hydrogenated coconut oil; fish oil feeding had no effect upon the IL-2 con-centration in the medium. In contrast, higher IL-2 production by Con A-stimulated spleen lymphocytes taken from autoimmune dis-ease-prone mice fed 100 g/kg fish oil com-pared with those fed 100 g/kg corn oil has been reported(76); the intracellular levels of mRNA for IL-2 were also elevated in the fish oil-fed mice although these levels could not be quantified. A recent study reported that inclusion of α-linolenic acid or eicosapen-taenoic plus docosahexaenoic acid in the diet of monkeys resulted in enhanced ex vivo production of IL-2 (77). The latter study related the discrepancies in the literature to the levels of vitamin E included in the diets used.
There have been few animal studies of the effects of dietary lipids on lymphocyte-derived cytokines other than IL-2. The stud-ies which have been reported suggest mini-mal effects of dietary fat upon production of IL-4, IL-10 and interferon (IFN)-γ (14,76), but more studies need to be performed to confirm this.
Human studies. Endres et al. (37) were
the first to show that supplementation of the diet of healthy volunteers with n-3 PUFAs diminishes ex vivo production of IL-1α, IL-1ß and TNF by human PBMNCs.
Interest-ingly, the production of these cytokines re-mained suppressed for 10 weeks after the supplementation had ended, indicating the long period of “washout” required to reverse the effects of fish oil supplementation. It is also worth noting that this study showed that the precise effects of dietary n-3 PUFAs vary according to the stimulus used to induce cytokine production. A later study by Endres et al. (25) using the same supplementation regime reported a decrease in ex vivo IL-2 production in response to Con A; the reduc-tion was greater 10 weeks after supplemen-tation had ended than at the end of supple-mentation. Virella et al. (78) reported that 6 weeks of supplementation of the diet with 2.4 g eicosapentaenoic acid per day resulted in lowered production of IL-2 by PBMNCs stimulated with PHA or PWM; in agreement with the long “washout” period suggested by Endres et al. (25,37), IL-2 production re-mained low 8 weeks after the supplementa-tion had ended but returned to pre-supple-mentation levels 22 weeks post-supplemen-tation. Meydani et al. (23) supplemented the diet of healthy young (20-33 years of age) and older (51-68 years of age) women with 2.4 g n-3 PUFAs per day and examined ex
vivo production of a range of cytokines after
4, 8 and 12 weeks. There were time-depend-ent decreases in the production of 1ß, IL-2, TNF and IL-6 by PBMNCs from the older women; production of TNF and IL-6 by cells from the young women was also significant-ly decreased and there were trends towards decreased production of IL-1ß and IL-2 by these cells. The same workers have studied the effect of including fish (providing 1.23 g
n-3 PUFAs/day) in a low fat, low cholesterol
diet for 24 weeks; the ex vivo production of IL-1ß, IL-6 and TNF by PBMNCs was sig-nificantly reduced and there was a nonsig-nificant reduction in IL-2 production (24). The production of IL-1 and IL-6 in whole blood cultures was decreased if the subjects had consumed 1.1 to 1.6 g n-3 PUFAs per day for 6 to 8 weeks but only if the cultures
were stimulated with “low” concentrations (≤0.001 µg/ml) of LPS (79). Interestingly, if higher LPS concentrations were used there was no effect of the supplementation upon production of these two cytokines; TNF production was unaffected by n-3 PUFAs in this study. Recently, the effects of di-etary α-linolenic acid upon IL-1ß and TNF-α production by human cells have been reported; subjects consumed a sunflower oil-rich diet (which was very similar to their typical diet) or a diet rich in α-linolenic acid which was provided by linseed oil capsules and linseed oil-based spreads and cooking oils (80). In this way the linseed oil consumption increased to a mean of 13.7 g/ day. Ex vivo production of both IL-1ß and TNF-α by PBMNCs was decreased by the linseed oil diet. If the subjects then supple-mented their diet with encapsulated eicosa-pentaenoic plus docosahexaenoic acids (2.7 g/day) production of both cytokines was further decreased. These authors showed a correlation between mononuclear cell eicosapentaenoic acid content and produc-tion of IL-1ß and TNF-α.
Supplementation of the diet of healthy volunteers or multiple sclerosis patients with encapsulated fish oil providing approximately 5 g n-3 PUFAs per day for 24 weeks resulted in lower ex vivo production of IL-1ß, TNF-α, IL-2 and IFN-γ by PBMNCs (81). An earlier study reported lower production of IL-1 by PBMNCs taken from rheumatoid arthritis patients who had consumed n-3 PUFAs for 24 weeks (82); interestingly, cells from the patients who consumed an olive oil placebo (9 g/day) in this trial also showed diminished IL-1 production. There was increased pro-duction of IL-2 by PBMNCs from rheuma-toid arthritis patients who supplemented their diet with olive oil or a “low” level of n-3 PUFAs for 24 weeks; this increase did not occur in patients consuming a “high” level of n-3 PUFAs (82).
Recently, circulating cytokine levels in patients receiving intravenous infusions of
lipid emulsions post-surgery were reported (83); patients received either a medium chain/ long chain triacylglycerol mix (50:50 v/v) or this mix also containing fish oil (50:30:20 v/ v/v). Patients received 50 g fat per day on days 1 and 2 post-surgery and 100 g fat per day on days 3, 4 and 5; thus patients in the group receiving the fish oil-containing emul-sion received approximately 3 (days 1 and 2) or 6 (days 3, 4 and 5) g n-3 PUFAs per day. Plasma TNF-α levels were significantly lower in the fish oil group at 6 days post-surgery; plasma IL-6 levels were lower (but not sig-nificantly) 10 days surgery and the post-surgery increase in plasma IL-10 levels was reduced in this group.
n-3 PUFAs and cytokine receptor expression Feeding rats a diet containing 200 g/kg fish oil lowered the proportion of spleen and thymic lymphocytes bearing the IL-2 recep-tor (IL-2R; CD25) following Con A stimula-tion (13). Spleen lymphocytes from rats fed fish oil also showed a lower level of expres-sion of the IL-2R following mitogenic stim-ulation (15). In accordance with these ani-mal studies, supplementation of the diet of patients with psoriasis or atopic dermatitis with n-3 PUFA ethyl esters (6 g/day) caused a significant reduction in the percentage of IL-2R+ blood lymphocytes following PHA stimulation(84); the level of expression of the IL-2R on the positive cells was also significantly reduced.
n-3 PUFAs and nitric oxide production by macrophages
Two studies using rat resident peritoneal macrophages reported that NO production is significantly diminished by fish oil feeding (68,85). In contrast, Hubbard et al. (35) found no effect of feeding mice 100 g/kg linseed or fish oil upon ex vivo NO production by thioglycollate-elicited macrophages; never-theless these workers, as well as Somers et
al. (32), reported that n-3 PUFA feeding diminishes macrophage cytotoxicity towards NO-sensitive P815 cells, suggestive of re-duced NO production. Yaqoob and Calder (53) found that several high fat diets, includ-ing 200 g/kg fish oil, enhanced NO produc-tion by murine thioglycollate-elicited mac-rophages compared with a low fat diet; there were no significant differences in NO pro-duction by macrophages from mice fed dif-ferent high fat diets. Fish oil has been re-ported to enhance NO production by murine peritoneal (34), rat lung (66) and pig lung (75) macrophages.
Recently, Harris et al. (86) reported in-creased appearance of NO metabolites in the urine of healthy human volunteers who supplemented their diet with 5 g fish oil concentrate per day (providing 3 g eicosa-pentaenoic acid plus docosahexaenoic acid) for 3 weeks; interestingly, 3 g of eicosapen-taenoic acid alone did not alter urinary NO metabolite levels. The levels measured by Harris et al. (86) were assumed to indicate whole body NO production; as such, a vari-ety of sources of NO are likely to exist and the authors concluded that much of it origi-nated from endothelial cells. There is much direct and circumstantial evidence that n-3 PUFAs enhance NO production by endothe-lial cells (see 86 for references).
n-3 PUFAs and adhesion molecule expression Adhesion molecules are involved in many cell-to-cell interactions. For example, inter-action between T lymphocytes and antigen-presenting cells is in part mediated by the ligand-receptor pairs CD11a/CD18:CD54, CD11a/CD18:CD102 and CD2:CD58 (for reviews, see 87-89). Thus, an efficient cell-mediated immune response requires appro-priate levels of expression of these mol-ecules on T lymphocytes. In addition, leuco-cyte adhesion to the endothelium involves a number of ligand-receptor pairs including CD11a/CD18:CD54, CD54:CD11a/CD18,
CD49d/CD29:CD106, CD2:CD58, CD62L: MAdCAM-1 and CD44: hyaluraonate (for reviews, see 87-89). Thus, the movement of leucocytes between body compartments, into and out of lymphoid organs and into sites of immune or inflammatory reactivity requires adhesion molecule expression. Adhesion molecule expression appears to be involved in several acute and chronic inflammatory disease processes (for a review, see 90), and antibodies against certain adhesion molecules can reduce chronic inflammatory disease (90).
In vitro studies of n-3 PUFAs and adhe-sion molecule expresadhe-sion. It has become
ap-parent that n-3 PUFAs can affect adhesion molecule expression by some cell types, at least in vitro. Calder et al. (91) observed that murine thioglycollate-elicited peritoneal mac-rophages cultured in the presence of eicosa-pentaenoic or docosahexaenoic acid were less adherent to artificial surfaces (the adhe-sion to one of these surfaces is mediated by CD11a/CD18) than those cultured with some other fatty acids. This observation suggests that n-3 PUFAs decrease expression of ei-ther CD11a or CD18 or both proteins. More recently, De Caterina et al. (92) reported that culture of human adult saphenous vein en-dothelial cells with docosahexaenoic acid significantly decreased the cytokine-induced expression of CD106, CD62E and CD54 in a dose-dependent manner. The adhesion of U937 monocytes or human peripheral blood monocytes to the endothelial cells was di-minished following incubation of the latter with docosahexaenoic acid (92). Since the binding between monocytes and endothelial cells partially depends upon CD106 expres-sion on the endothelial cells, the reduced expression of CD106 caused by docosahex-aenoic acid appears to have a functional effect. Kim et al. (93) reported that incuba-tion of LPS-stimulated pig aortic endothelial cells with eicosapentaenoic acid resulted in diminished binding between these cells and U937 monocytes. Eicosapentaenoic acid was
shown to reduce the expression of CD106, CD62E and CD54 on the surface of LPS-stimulated human umbilical vein endotheli-al cells (93). Recently, it was shown that inclusion of eicosapentaenoic or docosahex-aenoic acid in the medium of cultured rest-ing or LPS- or cytokine-stimulated human umbilical vein endothelial cells decreased the ability of peripheral blood lymphocytes to bind (94); both n-3 PUFAs were shown to decrease the level of expression of CD106, but not of CD54 or CD62E, on the surface of cytokine-stimulated endothelial cells. This latter study also reported, for the first time, the in vitro effect of n-3 PUFAs on adhesion molecule expression on lymphocytes: incu-bation of lymphocytes with either eicosa-pentaenoic or docosahexaenoic acid reduced the level of expression of CD11a and CD62L but did not affect CD49d expression (94). In parallel with this reduction, the binding of lymphocytes to untreated or cytokine-stimu-lated endothelial cells was diminished. In another recent study, incubation with eicosa-pentaenoic acid was shown to reduce the level of expression of CD54, but not CD11a, on the surface of resting or IFN-γ-stimulated human monocytes (95); docosahexaenoic acid did not alter expression of these mol-ecules. Thus, it appears that culture of macrophages, monocytes, lymphocytes or endothelial cells with n-3 PUFAs can de-crease adhesion molecule expression result-ing in diminished ability to bind to other cell types.
Effects of dietary n-3 PUFAs on adhe-sion molecule expresadhe-sion. There are few
stud-ies of the effects of inclusion of n-3 PUFAs in the diet upon adhesion molecule expres-sion, although it was recently shown that supplementation of the human diet with n-3 PUFAs results in significantly lower levels of expression of CD11a and CD54 on pe-ripheral blood monocytes (45). Feeding rats a diet containing 200 g fish oil/kg signifi-cantly reduced (by 20 to 35%) the levels of expression of CD2 and CD11a on freshly
prepared lymphocytes and of CD2, CD11a and CD54 on Con A-stimulated lympho-cytes (15). Furthermore, the levels of CD2, CD11a and CD54 were reduced on popliteal lymph node lymphocytes following localised graft vs host or host vs graft responses in vivo (96). Reduced adhesion molecule expres-sion suggests that cells will be less able to interact with receptor-bearing cells. In ac-cordance with this, we have recently served that lymph node lymphocytes ob-tained from fish oil-fed rats adhere less well to macrophage and endothelial cell mono-layers (Sanderson P and Calder PC, unpub-lished observations). These observations sug-gest that n-3 PUFA feeding will affect the movement of lymphocytes and monocytes between body compartments and perhaps into sites of inflammatory or autoimmune activity.
Dietary fatty acids and in vivo meas-ures of cell-mediated immunity and inflammation
The studies outlined above have investi-gated the effects of dietary manipulations upon ex vivo functions of isolated cell popu-lations. Although a number of consistent patterns have emerged from these studies, there are also contradictory reports and it is evident that the outcome of such ex vivo measures is strongly influenced by the ex-perimental conditions used. Furthermore, in
vivo cells exist as part of a network being
influenced by other cell types; often such interactions are disturbed by the purification of the particular cell types to be studied. Therefore, it is important to investigate the effect of dietary fats on the intact, fully functioning system in which all normal cel-lular interactions are in place. The ability to make in vivo measures of inflammation and cell-mediated immunity offers the prospect of investigating the effects of dietary ma-nipulations upon the overall responses of these systems.
Acute inflammatory responses
Arachidonic acid-derived eicosanoids are involved in mediating inflammatory re-sponses. Since n-3 PUFAs diminish the pro-duction of these mediators, they should exert anti-inflammatory activities. Indeed, n-3 PUFAs decrease the production of pro-in-flammatory mediators such as PGE2 and LTB4 during antigen-induced inflammation of the air pounch (97), carrageenan-induced inflammation of the footpad (97) and zymo-san-induced peritoneal inflammation (98). The latter study also reported that dietary n-3 PUFAs inhibited the influx of neutrophils into the peritoneal cavity which accompa-nies such treatment. Feeding rats 100 g/kg cod liver oil for 10 weeks significantly low-ered (by 40%) the inflammatory response to carrageenan injection into the footpad com-pared with feeding coconut oil or groundnut oil (99). In accordance with that observa-tion, feeding rats high fat diets containing ethyl esters of eicosapentaenoic or docosa-hexaenoic acid resulted in a 50% reduction in footpad swelling in response to carrageen-an injection compared with feeding safflower oil (100); both n-3 PUFAs were equally effective.
In vivo responses to endotoxin and cytokines Intravenous infusions of a 10% (v/v) li-pid emulsion rich in fish oil into guinea pigs significantly enhanced survival to intraperi-toneally injected LPS compared with infu-sion of a 10% (v/v) safflower oil emulinfu-sion (101); the total amount of lipid infused was 13 g/animal. The same workers later showed that feeding a fish oil-rich diet to guinea pigs for 6 weeks significantly increased survival to an intraperitoneal injection of LPS com-pared with animals fed a safflower oil-rich diet (102).
In accordance with the diminished sus-ceptibility to the lethal effects of endotoxin in experimental animals, feeding weanling
rats a 100 g/kg fish oil diet for 8 weeks significantly decreased a number of re-sponses to intraperitoneal TNF-α: the rises in liver zinc and plasma C3 concentrations, the fall in plasma albumin concentration and the increases in liver, kidney and lung protein synthesis rates were all prevented by the fish oil diet (103). Fish oil feeding to rats or guinea pigs also diminishes the pyro-genic (104,105) and anorexic effects (103, 106) of IL-1 and TNF-α compared with feeding n-6 PUFA-containing oils.
Delayed-type hypersensitivity
The delayed-type hypersensitivity (DTH) reaction is the result of a cell-medi-ated response to challenge with an antigen to which the individual has already been primed. There are several reports that the DTH response in rodents is significantly reduced by n-3 PUFAs. For example, intra-gastric administration of a fish oil concen-trate to rats for 50 days lowered the DTH response to bovine serum albumin com-pared with administering safflower oil, ol-ive oil or water(97) while addition of ethyl esters of either eicosapentaenoic or docosa-hexaenoic acid to the diet of mice consum-ing a safflower oil diet reduced the DTH response to tuberculin (107); both n-3 PUFAs were equally effective. The DTH response to sheep red blood cells in mice was diminished following tail-vein injec-tions of emulsions of triglycerides rich in eicosapentaenoic or docosahexaenoic acid (108); a soybean oil emulsion was injected in the control animals. Most recently, it was reported that feeding aged beagle dogs a long-chain n-3 PUFA-rich diet significant-ly diminished the DTH response to intra-dermal KLH compared with feeding diets containing smaller amounts of long-chain
n-3 PUFAs (109).
Feeding a linseed oil-rich diet to healthy human volunteers for 8 weeks lowered the DTH response to 7 recall antigens, although
this reduction was not statistically signifi-cant (110). Supplementation of the diet of volunteers consuming a low fat, low cho-lesterol diet with 1.25 g n-3 PUFA/day di-minished the DTH responses to 7 recall antigens (24).
Graft vs host and host vs graft responses The so-called popliteal lymph node (PLN) assay provides a useful experimental model in rodents for measuring graft vs host (GvH) and host vs graft (HvG) re-sponses elicited by injection of allogeneic cells into the footpad of the host. The GvH response primarily involves the polyclonal activation, and subsequent proliferation, of host B-cells, although NK cells may also be involved in the host defence. In contrast, the HvG reaction is a T-cell-mediated re-sponse, in which CTLs of the host recognise MHC antigens on the injected cells. In both cases the enlargement in PLN size (more than 15-fold in the GvH response and 4-fold in the HvG response) is due largely to proliferation of activated host cells; most of these originate within the PLN although there is also some recruitment of cells from the bloodstream. A suppressed HvG re-sponse was observed in mice fed a 160 g/kg fish oil diet compared with those fed a stan-dard chow diet (111); lower levels of fish oil (25, 50, 100 g/kg) did not significantly affect the response. Significantly dimin-ished GvH and HvG responses (by 34% and 20%, respectively) were observed in rats fed 200 g/kg fish oil compared with those fed a low fat diet or diets containing 200 g/kg coconut, olive, safflower or evening primrose oils (96). Such observa-tions accord with the demonstraobserva-tions of sig-nificantly diminished ex vivo T lymphocyte proliferation, NK cell activity and CTL ac-tivity following fish oil feeding. Feeding rats a 200 g/kg fish oil diet resulted in fewer IL-2R+ cells and CD16+/CD3- cells in the popliteal lymph node following the GvH
response, indicating an inhibition of lym-phocyte activation and a decrease in the proportion of NK cells, respectively (96). Recently, a dose-dependent effect of lin-seed oil compared with sunflower oil upon the GvH response in rats was reported (18). Animal models of organ transplantation
Graft rejection in transplantation sur-gery is caused by an immune reaction to the foreign material introduced into the body; T-cells have been implicated in ac-celerated graft rejection, but antibodies with specificity for the graft donor have also been observed following rejection, implying that both cell-mediated immuni-ty and humoral immuniimmuni-ty play a part in the rejection process. Studies investigating the effects of eicosanoids on organ transplan-tation pre-date those investigating the ef-fects of fatty acids. Some eicosanoids pro-mote graft rejection while others propro-mote graft survival (for a review, see 112). Therefore, since n-6 and n-3 PUFAs will affect the levels and types of eicosanoids formed, they would be expected to influ-ence graft survival. In addition, PUFAs exert immunomodulatory effects which might be independent of eicosanoids and so these effects might play a role in en-hancing graft survival. The effect of feed-ing n-3 PUFAs to recipient or donor or both upon subsequent cardiac allograft sur-vival was investigated in rats (113). It was found that survival was prolonged if the recipients were fed fish oil or ethyl esters of eicosapentaenoic or docosahexaenoic acid regardless of the diet of the donor, or if the donors were fed fish oil or ethyl esters of eicosapentaenoic or docosa-hexaenoic acid or fish oil regardless of the diet of the recipient or if both donor and recipient were fed fish oil. Greater prolon-gation of cardiac survival in rats receiving an infusion of fish oil post-transplantation compared with those receiving soybean
oil infusion was reported (114); in turn soybean oil enhanced survival compared with saline infusion. Oral fish oil (4.5 g/ day) has also been shown to prolong the survival of islets of Langerhans grafts in mice (115).
Effects of n-3 PUFAs in animal models of inflammatory and autoimmune diseases
Dietary fish oil has been shown to have significantly beneficial clinical, immuno-logical and biochemical effects in a num-ber of animal disease models. These ef-fects include increased survival and de-creased proteinuria and anti-DNA antibod-ies in mice with autoimmune glomerulo-nephritis, a model for systemic lupus erythematosus (41,116-119), decreased joint inflammation in rodents with col-lagen-induced arthritis (120,121) and less inflammation in rats with various models of colitis (122-125). These observations suggest that diets enriched in n-3 PUFAs might be of some therapeutic benefit in these diseases in man.
Clinical studies of the effects of dietary fatty acids in inflammatory and autoimmune diseases and in transplantation
Inflammatory and autoimmune diseases The low incidence of cardiovascular disorders amongst populations consuming large quantities of oily fish has been well documented (126-128), but the intense in-terest in fish oil and heart disease has overshadowed the unusual pattern of the incidence of some other diseases in native Greenland Eskimos. Kromann and Green (127) described a very low incidence of autoimmune and inflammatory disorders, such as psoriasis, asthma and type-I diabe-tes, and the complete absence of multiple sclerosis in a population of Greenland
Es-kimos compared with sex- and age-matched groups living in Denmark. Thus, the n-3 PUFA-containing fish oil in the Eskimo diet could have a protective role towards these types of diseases. Most of these diseases are characterised by inap-propriate activation of T-cells resulting in attack on, and ultimately destruction of, host tissues. Typically, the sites of tissue destruction (e.g., joints in rheumatoid ar-thritis, neural tissue in multiple sclerosis) contain activated T-cells and macrophages and mediators produced by these cells, such as cytokines, eicosanoids and reac-tive oxygen species. The favourable out-come resulting from dietary fish oil in animal models of inflammatory and au-toimmune diseases (see above) indicates that there may be some benefit from sup-plementation of the diet of suitable pa-tients with n-3 PUFAs. There have been a number of clinical trials assessing the ben-efits of dietary supplementation with fish and other oils in several inflammatory and autoimmune diseases; these studies have been reviewed elsewhere (7,129-134). Transplantation
The animal studies described above indi-cate that PUFAs, particularly n-3 PUFAs, could be used to prolong the survival of organ transplants. Recipients of kidney trans-plants who received 9 g fish oil/day for one year post-transplant (in conjunction with cyclosporin A and prednisolone) had signif-icantly improved glomerular filtration rate and significantly diminished cyclosporin A nephrotoxicity although there was no effect upon graft survival (135). A similar finding was made by Homan van der Heide et al. (136) who reported that renal transplant pa-tients who received fish oil (6 g/day for the first post-operative year) in combination with cyclosporin A had better kidney function and less rejections over one year compared with patients who received coconut oil and
cyclosporin A. Better kidney function in kid-ney graft recipients who consumed fish oil (8 g/day for one year post-transplant) was reported, although there was no reduction in rejection episodes compared with controls (137). Bennett et al. (138) reported no rejec-tion incidents in a group of kidney transplant recipients who received 9 or 18 g eicosapen-taenoic acid/day for 16 weeks post-trans-plantation; cyclosporin nephrotoxicity did not occur in this group but did occur in the control group who supplemented their diet with corn oil.
Mechanisms by which n-3 PUFAs might exert their effects
Clearly n-3 PUFAs do influence the func-tional activities of cells of the immune sys-tem, although a number of conflicting obser-vations have been made. These fatty acids appear to alter the production of mediators involved in communication between cells of the immune system (eicosanoids, cytokines, NO) and to alter the expression of key cell surface molecules involved in direct cell-to-cell contact (adhesion molecules). The pro-duction of cytokines and NO is regulated by eicosanoids and so an n-3 PUFA-induced change in the amount and types of eicosanoids formed could, at least partially, explain the effects of n-3 PUFAs. However, many of the effects of n-3 PUFAs appear to be exerted in an eicosanoid-independent manner. Thus, other mechanisms of action of n-3 PUFAs upon immune cell function must be consid-ered. One such mechanism could be through regulating expression of key genes involved in immune cell functioning and in the pro-duction of immune cell-derived mediators. Evidence that n-3 PUFAs affect gene expres-sion in cells of the immune system
It is now well documented that fatty acids affect the expression of genes involved in hepatic fatty acid and lipoprotein
metabo-lism (for reviews, see 139-141) and the genes involved in adipocyte differentiation and development (for reviews, see 140,142). Di-etary n-3 PUFAs have particularly potent effects upon the expression of genes for proteins involved in hepatic peroxisomal proliferation, fatty acid oxidation and lipo-protein assembly (e.g., 143). Studies of the effects of fatty acids upon the expression of genes important in immune cell functioning are few and relatively recent.
n-3 PUFAs and cytokine gene expres-sion. Inclusion of fish oil in the diet of
au-toimmune disease-prone mice resulted in elevated levels of mRNA for IL-2, IL-4 and TGF-ß and reduced levels of mRNA for c-myc and c-ras in the spleen (76). The same authors also showed that dietary fish oil completely abolished mRNA production for IL-1ß, IL-6 and TNF-α in the kidneys of these animals (144). It has been reported that feeding mice a fish oil-rich diet significantly diminished the level of IL-1ß mRNA in LPS-or phLPS-orbol ester-stimulated spleen lympho-cytes(145); the lower IL-1ß mRNA level was not due to accelerated degradation but to impaired synthesis. Fish oil feeding to mice lowered basal and LPS-stimulated TNF-α mRNA levels in peritoneal macrophages (34).
n-3 PUFAs and adhesion molecule gene expression. De Caterina et al. (92) showed
that incubation of human saphenous vein endothelial cells with docosahexaenoic acid results in reduced levels of mRNA for CD106.
n-3 PUFAs and NO synthase gene ex-pression. Khair-El-Din et al. (146) reported
that incubation of murine thioglycollate-elic-ited peritoneal macrophages with docosa-hexaenoic acid decreased NO production in response to LPS plus IFN-γ (or IFN-γ plus TNF-α) and that the inhibition was concen-tration dependent. Neither arachidonic nor eicosapentaenoic acids at concentrations of up to 100 µM affected NO production. This study found that incubation of the cells with docosahexaenoic acid resulted in a lower
level of mRNA for inducible NO synthase (iNOS), the enzyme responsible for NO pro-duction in macrophages stimulated in this way; the lower level of iNOS mRNA was due to an inhibition of transcription (146). Mechanisms by which n-3 PUFAs might affect gene expression
Thus, it is now apparent that n-3 PUFAs might affect immune cell function partly by regulating the expression of genes encoding for proteins involved in cellular responses and in communication between cells. One mechanism by which n-3 PUFAs could af-fect gene expression is through changes in the signal transduction pathways which link cell surface receptors to the activation of nuclear transcription factors. Alternatively,
n-3 PUFAs (indeed PUFAs in general) or
their derivatives might bind directly to nuclear transcription factors thereby altering their activity.
Fatty acids and signal transduction.
Many lipids are involved directly in intracel-lular signalling pathways; for example, hy-drolysis of membrane phospholipids such as PIP2 and phosphatidylcholine by phospholi-pases generates second messengers such as DAG and inositol-1,4,5-trisphosphate (IP3). Other phospholipids play a role in activating or stabilising enzymes involved in intracel-lular signalling; for example, phosphatidyl-serine is required for protein kinase C (PKC) activation. DAG activates some isoforms of PKC and also activates sphingomyelinase to release the second messenger ceramide. Ceramide in turn may be converted to osine and sphingosine-1-phosphate; sphing-osine inhibits PKC and activates some phos-pholipases. Since PIP2, phosphatidylcholine, phosphatidylserine and DAG all contain fatty acyl chains attached to the sn-1 and -2 posi-tions of the glycerol moiety, it is conceivable that changing the type of fatty acid present may alter the precise properties of these compounds with regard to their functions in
signal transduction. Certainly, changing the fatty acid composition of the diet (e.g., by feeding fish oil) markedly alters the phos-pholipid and DAG molecular species com-positions of lymphocytes (147) and macro-phages (148). That such changes might di-rectly influence signal transduction pathways is shown by the observations that PKC is more active in the presence of dioleoylglyc-erol or diarachidonoylglycdioleoylglyc-erol than in the presence of DAGs containing two saturated fatty acids or one saturated and one unsatur-ated fatty acid (149) and that rat spleen PKC is less active in the presence of an n-3 PUFA-rich phosphatidylserine compared with a PUFA-poor phosphatidylserine (150).
In support of the idea that n-3 PUFAs influence intracellular signalling pathways, DAG generation was reduced in Con A-stimulated lymphocytes taken from mice fed the ethyl ester of docosahexaenoic acid com-pared with those taken from safflower oil-fed mice (151). More recently, it was shown that feeding either eicosapentaenoic or do-cosahexaenoic acid to mice results in re-duced DAG generation by Con A-stimulated spleen lymphocytes (16); this study also showed, for the first time, that fish oil-de-rived n-3 PUFAs suppress ceramide genera-tion in Con A-stimulated lymphocytes. The DAG in these studies could have been gener-ated by the activity of phosphatidylcholine PLC, phospholipase D or phosphatidylinosi-tol PLC or a combination of these. A recent study which showed reduced calcium iono-phore-stimulated DAG production in mac-rophages from eicosapentaenoic acid-fed mice compared with those fed an n-6 PUFA-rich diet suggested that phosphatidylcholine hydrolysis contributed significantly to the total DAG formed (148), implicating phos-phatidylcholine PLC and/or phospholipase D in DAG formation. These data indicate that n-3 PUFAs in some way affect the activ-ity of one or more phospholipases respon-sible for the generation of key second mes-sengers; this may be via changes in the fatty
acid composition of the substrate phospho-lipids. Addition of 14.4 g per day of fish oil-derived n-3 PUFAs to the diet for 10 weeks resulted in lower platelet activating factor-or LTB4-stimulated IP3 generation in periph-eral blood neutrophils (40). The platelet ac-tivating factor and LTB4 receptors are coupled via G-proteins to PLCß. In a recent study, it was observed that calcium iono-phore- or Con A-stimulated IP3 generation in rat lymphocytes was significantly reduced if the lymphocytes came from animals fed fish oil (Sanderson P and Calder PC, unpub-lished observations). Furthermore, it was found that, although the amount of PLCγ-1 in rat lymphocytes was unaffected by diet, the ability to phosphorylate, and so to acti-vate, the enzyme in response to suitable stimuli appeared to be significantly impaired by fish oil feeding (Sanderson P and Calder PC, unpublished observations). Lymphocyte PLCγ-1 is activated by one or more tyrosine kinases (lck, fyn, ZAP-70); thus, these ob-servations are suggestive of lowered activity of certain tyrosine kinases following fish oil feeding. How n-3 PUFAs inhibit tyrosine kinase activity is not clear but these tyrosine kinases are associated with specific regions of the plasma membrane. This association might require certain phospholipid fatty acid compositions or membrane physical proper-ties which could be markedly altered by n-3 PUFA incorporation, thus resulting in an inability of the tyrosine kinase to maintain its optimal activity.
It has been proposed that unsaturated fatty acids themselves may have a direct effect upon intracellular signalling pathways (for a review, see 152). This direct modula-tory effect of fatty acids has been most ex-tensively documented in relation to PKC activity which was shown to be enhanced by docosahexaenoic acid (153). In contrast, al-though it was reported that eicosapentaenoic and docosahexaenoic acids increased brain PKC activity in the absence of phosphatidyl-serine and DAG, in the presence of
phospha-tidylserine and DAG both n-3 PUFAs caused up to 60% inhibition of PKC activity (154). Other studies have shown that eicosapen-taenoic and docosahexaenoic acids inhibit rat lymphocyte or macrophage PKC activity in the presence of calcium, phosphatidylser-ine and DAG(155,156), whereas protein kinase A activity was unaffected by these fatty acids. In accordance with both the di-rect effects of n-3 PUFAs in vitro (155,156) and the effects of enriching phosphatidylser-ine with n-3 PUFAs (150), feeding mice fish oil resulted in diminished spleen lympho-cyte PKC activity (157).
A change in the concentration of intra-cellular free calcium is often a key compo-nent in the intracellular signalling pathway which follows the stimulation of lympho-cytes, macrophages and other cells by growth factors, cytokines and antigens. There is now considerable evidence that free fatty acids influence these changes. For example, it was reported that oleic acid, but not stearic acid, inhibited the target cell- or Con A-stimulated rise in intracellular free calcium concentra-tion in a cytotoxic T-cell line (158). Several unsaturated fatty acids including α-linolenic, eicosapentaenoic and docosahexaenoic ac-ids inhibit the anti-CD3-induced increase in intracellular free calcium concentration in the Jurkat T-cell line(159,160); the fatty acids appeared to act by blocking calcium entry into the cells and it was concluded that they act directly upon receptor-operated cal-cium channels.Reduced IP3 levels as a result of diminished phospholipase activity (see above) would also contribute to decreased intracellular free calcium concentrations, which in turn would reduce the activity of some isoforms of PKC.
Fatty acids and transcription factors within cells of the immune system.
Lympho-cytes and other immune and inflammatory cells contain many transcription factors in-cluding nuclear transcription factor-κB (NF-κB), nuclear transcription factor of activated T-cells (NFAT), AP-1, various oncogene
products (e.g., myc, fos, jun), steroid hor-mone receptors and specific transcription factors such as IL-2, IL-6 and NF-ICAM-1. NF-κB plays a key role in inducing the production of many key mediators within the immune system: NF-κB regulates the synthesis of cytokines including IL-1, IL-2, IL-6, TNF-α and IFN-γ, of cytokine recep-tors including IL-2R, of adhesion molecules including CD54, CD62E and CD106, of en-zymes involved in mediator generation such an iNOS and of a range of acute phase proteins (for a review, see 161). NF-κB is activated by the phosphorylation and subse-quent dissociation of one of its three sub-units, the so-called inhibitory-κB (IκB); this leaves the remaining dimer free to translo-cate to the nucleus and bind to appropriate response elements on target genes. It ap-pears that, at least in response to some stimuli, the phosphorylation of IκB is performed by PKC. Given the effects of n-3 PUFAs out-lined previously (e.g., reduced PLC activity, resulting in reduced DAG and IP3 genera-tion, resulting in turn in a reduced intracellu-lar free calcium rise and a reduced activation of PKC isoforms), it is evident how these fatty acids could prevent activation of NF-κB and so suppress expression of a range of genes including those for cytokines, cytokine receptors, adhesion molecules and iNOS. Ceramide can also activate NF-κB independ-ently of PKC activity; thus, the observation of Jolly et al. (16) of reduced ceramide pro-duction within lymphocytes from n-3 PUFA-fed mice could also partly account for the reduced functional responses (e.g., IL-2 pro-duction and proliferation) of these cells.
Some transcription factors are receptors for lipophilic molecules; these include ste-roid hormone receptors and the family of receptors known as peroxisomal proliferator activated receptors (PPARs). Eicosanoid derivatives of arachidonic acid including PGA2, PGB2, PGD2, 15-deoxy-∆12,14-PGJ2, but not PGE2, PGF2α or PGI2, are activators of PPARα, γ and δ (162); LTB4, 8-HETE
and 12-HETE appear to activate PPARα only (162). In contrast to their role as PPAR activators, most eicosanoids are poor PPAR ligands (162), although 8-HETE is a high affinity ligand for PPARα (162). The claim that LTB4 is a PPARα ligand (163) is now disputed (162). It has been claimed that 15-deoxy-∆12,14-PGJ
2 is a ligand for PPARγ (164-166) and PPARα (166), although Forman et al. (162) showed it does not bind PPARα or δ. Interestingly, the eicosapentaenoic acid derivative 8-HEPE is a potent activator of, and ligand for, PPARα (162). The various roles of arachidonic and eicosapentaenoic acid derivatives as activators and ligands of PPAR isoforms provide a mechanism by which these fatty acids can affect the activity of these transcription factors. It has been known for some time that PUFAs
them-selves are activators of PPARs (for reviews, see 140,141). Very recently, and in contrast to earlier reports, it has been found that unsaturated fatty acids (linoleic, α-linolenic, γ-linolenic, arachidonic, eicosapentaenoic, and docosahexaenoic acids) bind directly to PPARα (162,166), δ (162) and γ (166). Thus, it appears that PUFAs are both activators of and ligands for some forms of PPARs. The mechanism of action and functional effects of PPARs have been reviewed in detail re-cently (140,141). PPAR isoforms have been identified in lymphoid tissues. Using in situ DNA hybridisation and immunohistochemi-cal staining Braissant et al. (167) identified moderate to very strong expression of PPARα, ß and γ in rat spleen (both red and white pulp) and lymph nodes; Kliewer et al. (168) had earlier reported the presence of
Figure 3 - Schematic representa-tion of the mechanisms by which n-3 PUFAs might influ-ence immune cell function.
PPARγ and δ mRNA in murine spleen. These observations suggest that genes within cells of the immune system will be subject to regulation by ligands for PPAR isoforms; these include PUFAs and PUFA derivatives. Cells of the immune system also possess steroid hormone, vitamin D and retinoic acid receptors making them subject to the effects of ligands of these transcription factors. In-cidentally, it is proposed that PUFAs regu-late the pathway of activation of steroid hor-mone receptors (for reviews, see 169-171), which may account for the observations that PUFAs, particularly n-3 PUFAs, sensitise
immune cells to the effects of steroid hor-mones (e.g., 172).
In addition to the effects which influence the activity of transcription factors, n-3 PUFAs might also regulate the synthesis of transcription factors. Fernandes et al. (76) reported markedly decreased c-myc mRNA in spleens from fish oil-fed mice; c-myc mRNA encodes a transcription factor.
Figure 3 attempts to give an overview of the sites at which n-3 PUFAs might influ-ence the functional activities of cells of the immune system.
References
1. Kinsella JE, Lokesh B, Broughton S & Whelan J (1990). Dietary polyunsaturated fatty acids and eicosanoids: potential ef-fects on the modulation of inflammatory and immune cells: An overview. Nutri-tion, 6: 24-44.
2. Yaqoob P & Calder PC (1993). The effects of fatty acids on lymphocyte functions. International Journal of Biochemistry, 25: 1705-1714.
3. Peck MD (1994). Interactions of lipids with immune function. II: Experimental and clinical studies of lipids and immunity. Journal of Nutritional Biochemistry, 5: 514-521.
4. Calder PC (1995). Fatty acids, dietary lip-ids and lymphocyte functions. Biochemi-cal Society Transactions, 23: 302-309. 5. Calder PC (1996). Can n-3 polyunsaturated
fatty acids be used as immunomodulatory agents? Biochemical Society Transac-tions, 24: 211-220.
6. Calder PC (1996). Effects of fatty acids and dietary lipids on cells of the immune system. Proceedings of the Nutrition So-ciety, 55: 127-150.
7. Calder PC (1996). Immunomodulatory and anti-inflammatory effects of n-3 polyun-saturated fatty acids. Proceedings of the Nutrition Society, 55: 737-774.
8. Calder PC (1997). Dietary fatty acids and the immune system. Nutrition Reviews (in press).
9. Miles EA & Calder PC (1997). Modulation of immune function by dietary fatty acids. Proceedings of the Nutrition Society (in press).
10. Alexander NJ & Smythe NL (1988).
Di-etary modulation of in vitro lymphocyte function. Annals of Nutrition and Metabo-lism, 32: 192-199.
11. Kelley DS, Nelson GJ, Serrato CM, Schmidt PC & Branch LB (1988). Effects of type of dietary fat on indices of im-mune status of rabbits. Journal of Nutri-tion, 118: 1376-1384.
12. Fritsche KL, Cassity NA & Huang S-C (1991). Effect of dietary fat source on anti-body production and lymphocyte prolif-eration in chickens. Poultry Science, 70: 611-617.
13. Yaqoob P, Newsholme EA & Calder PC (1994). The effect of dietary lipid manipu-lation on rat lymphocyte subsets and pro-liferation. Immunology, 82: 603-610. 14. Yaqoob P & Calder PC (1995). The effects
of dietary lipid manipulation on the pro-duction of murine T-cell-derived cyto-kines. Cytokine, 7: 548-553.
15. Sanderson P, Yaqoob P & Calder PC (1995). Effects of dietary lipid manipula-tion upon rat spleen lymphocyte funcmanipula-tions and the expression of lymphocyte surface molecules. Journal of Nutritional and En-vironmental Medicine, 5: 119-132. 16. Jolly CA, Jiang Y-H, Chapkin RS &
McMurray DN (1997). Dietary (n-3) poly-unsaturated fatty acids suppress murine lymphoproliferation, interleukin-2 secre-tion and the formasecre-tion of diacylglycerol and ceramide. Journal of Nutrition, 127: 37-43.
17. Marshall LA & Johnston PV (1985). The influence of dietary essential fatty acids on rat immunocompetent cell prostaglan-din synthesis and mitogen-induced
blas-togenesis. Journal of Nutrition, 115: 1572-1580.
18. Jeffery NM, Sanderson P, Sherrington EJ, Newsholme EA & Calder PC (1996). The ratio of n-6 to n-3 polyunsaturated fatty acids in the rat diet alters serum lipid lev-els and lymphocyte functions. Lipids, 31: 737-745.
19. Jeffery NM, Newsholme EA & Calder PC (1997). The level of polyunsaturated fatty acids and the n-6 to n-3 polyunsaturated fatty acid ratio in the rat diet both affect serum lipid levels and lymphocyte func-tions. Prostaglandins, Leukotrienes and Essential Fatty Acids, 57: 149-160. 20. Cathcart ES, Leslie CA, Meydani SN &
Hayes KC (1987). A fish oil diet retards experimental amyloidosis, modulates lym-phocyte function and decreases macro-phage arachidonate metabolism in mice. Journal of Immunology, 139: 1850-1854. 21. Berger A, German JB, Chiang BL, Ansari
AA, Keen CL, Fletcher MP & Gershwin MR (1993). Influence of feeding unsatur-ated fats on growth and immune status of mice. Journal of Nutrition, 123: 225-233. 22. Yaqoob P, Newsholme EA & Calder PC
(1995). Influence of cell culture conditions on diet-induced changes in lymphocyte fatty acid composition. Biochimica et Bio-physica Acta, 1255: 333-340.
23. Meydani SN, Endres S, Woods MM, Goldin BR, Soo C, Morrill-Labrode A, Dinarello C & Gorbach SL (1991). Oral (n-3) fatty acid supplementation suppresses cytokine production and lymphocyte pro-liferation: comparison between young and older women. Journal of Nutrition, 121:
